It is well-recognized that physical activity declines during adolescence and into early adulthood, a decline that is particularly dramatic among girls. While the majority of girls' physical activity declines, some maintain and others actually increase their physical activity levels throughout adolescence and into the transition to adulthood. These patterns are not well-characterized, although they can provide important clues as to how to reverse the physical activity decline. The girls who counter the trend of their peers' declining physical activity may have unique predictors of physical activity. The same may be true for sedentary behavior. It is now widely- recognized that physical activity and sedentary behaviors are two distinct classes of behavior. We propose to examine predictors of physical activity and excessive sedentary behavior as girl's transition from adolescence into emerging adulthood. We assembled a cohort of 589 girls who had 6 complete days of accelerometer to measure physical activity, along with extensive individual, social, school, and neighborhood variables assessed when the girls were in 8th and 11th grade. We propose to re-measure these girls when they will be approximately 22 years of age. With this additional measurement, we will be able to examine multi-level predictors of the trajectories of physical activity maintainers, increasers, and decliners, as well as excessive sedentary behavior maintainers, increasers and decliners. We can also assess how these trajectories influence weight status. Our study aims are 1) To examine multi-level predictors of the physical activity trajectories of a cohort of 589 adolescent girls from middle school (age 14) to high school (age 17) to emerging adulthood (age 22); 2) To examine multi-level predictors of the sedentary time trajectories of the cohort, and 3) to examine how physical activity and sedentary time trajectories predict weight status at age 22. We hypothesize that predictors of those who remain or become highly physically active will be different from those who remain or become inactive, and the predictors of those who remain or become highly sedentary will be different from those who remain or become less sedentary. We also hypothesize that those who remain or become physically active will have taken on fewer adult social roles and that those who remain or become sedentary will have taken on more adult social roles. Our ultimate goal for this body of research is to identify predictors amenable to change in order to develop effective interventions to stop the physical activity decline and rise o excessive sedentary behavior that occurs during adolescence and into young adulthood. Without truly understanding how these behaviors are occurring, the burden of chronic diseases will continue.